Recombinant DNA techniques have been used for production of heterologous proteins in transformed host cells. Generally, the produced proteins are composed of a single amino acid chain or two chains cleaved from a single polypeptide chain. More recently, multichain proteins such as antibodies have been produced by transforming a single host cell with DNA sequences encoding each of the polypeptide chains and expressing the polypeptide chains in the transformed host cell (U.S. Pat. No. 4,816,397).
The basic immunoglobulin (Ig) structural unit in vertebrate systems is composed of two identical "light" polypeptide chains (approximately 23 kDa), and two identical "heavy" chains (approximately 53 to 70 kDa). The four chains are joined by disulfide bonds in a "Y" configuration, and the "tail" portions of the two heavy chains are bound by covalent disulfide linkages when the immunoglobulins are generated either by hybridomas or by B cells.
A schematic of the general antibody structure is shown in FIG. 1. The light and heavy chains are each composed of a variable region at the N-terminal end, and a constant region at the C-terminal end. In the light chain, the variable region (termed "V.sub.L J.sub.L ") is the product of the recombination of a V.sub.L gene to a J.sub.L gene. In the heavy chain, the variable region (V.sub.H D.sub.H J.sub.H) is the product of recombination of first a D.sub.H and a J.sub.H gene, followed by a D.sub.H J.sub.H to V.sub.H recombination. The V.sub.L J.sub.L and V.sub.H D.sub.H J.sub.H regions of the light and heavy chains, respectively, are associated at the tips of the Y to form the antibody's antigen binding domain and together determine antigen binding specificity.
The (C.sub.H) region defines the antibody's isotype, i.e., its class or subclass. Antibodies of different isotypes differ significantly in their effector functions, such as the ability to activate complement, bind to specific receptors (Fc receptors) present on a wide variety of cell types, cross mucosal and placental barriers, and form polymers of the basic four-chain IgG molecule.
Antibodies are categorized into "classes" according to the C.sub.H type utilized in the immunoglobulin molecule (IgM, IgG, IgD, IgE, or IgA). There are at least five types of C.sub.H genes (C.mu., C.gamma., C.delta., C.epsilon., and C.alpha.), and some species (including humans) have multiple C.sub.H subtypes (e.g., C.gamma..sub.1, C.gamma..sub.2, C.gamma..sub.3, and C.gamma..sub.4 in humans). There are a total of nine C.sub.H genes in the haploid genome of humans, eight in mouse and rat, and several fewer in many other species. In contrast, there are normally only two types of light chain constant regions (C.sub.L), kappa (.kappa.) and lambda (.lambda.), and only one of these constant regions is present in a single light chain protein (i.e., there is only one possible light chain constant region for every V.sub.L J.sub.L produced). Each heavy chain class can be associated with either of the light chain classes (e.g., a C.sub.H.gamma. region can be present in the same antibody as either a .kappa. or .lambda. light chain).
A process for the immortalization of B cell clones producing antibodies of a single specificity has been developed involving fusing B cells from the spleen of an immunized mouse with immortal myeloma cells. Single clones of fused cells secreting the desired antibody could then be isolated by drug selection followed by immunoassay. These cells were given the name "hybridoma" and their antibody products termed "monoclonal antibodies."
The use of monoclonal antibodies as therapeutic agents for human disease requires the ability to produce large quantities of the desired antibody. One approach to increased production was simply to scale up the culture of hybridoma cells. Although this approach is useful, it is limited to production of that antibody originally isolated from the mouse. In the case where a hybridoma cell produces a high affinity monoclonal antibody with the desired biological activity, but has a low production rate, the gene encoding the antibody can be isolated and transferred to a different cell with a high production rate.
In some cases it is desirable to retain the specificity of the original monoclonal antibody while altering some of its other properties. For example, a problem with using murine antibodies directly for human therapy is that antibodies produced in murine systems may be recognized as "foreign" proteins by the human immune system, eliciting a response against the antibodies. A human anti-murine antibody (HAMA) response results in antibody neutralization and clearance and/or potentially serious side-effects associated with the anti-antibody immune response. Such murine-derived antibodies thus have limited therapeutic value.
One approach to reducing the immunogenicity of murine antibodies is to replace the constant domains of the heavy and light chains with the corresponding human constant domains, thus generating human-murine chimeric antibodies. Chimeric antibodies are generally produced by cloning the antibody variable regions and/or constant regions, combining the cloned sequences into a single construct encoding all or a portion of a functional chimeric antibody having the desired variable and constant regions, introducing the construct into a cell capable of expressing antibodies, and selecting cells that stably express the chimeric antibody. Examples of methods using recombinant DNA techniques to produce chimeric antibodies are described in PCT Publication No. WO 86/01533 (Neuberger et al.), and in U.S. Pat. No. 4,816,567 (Cabilly et al.) and U.S. Pat. No. 5,202,238 (Fell et al.).
In another approach, complementarity determining region (CDR)-grafted humanized antibodies have been constructed by transplanting the antigen binding site, rather than the entire variable domain, from a rodent antibody into a human antibody. Transplantation of the hypervariable regions of an antigen-specific mouse antibody into a human heavy chain gene has been shown to result in an antibody retaining antigen-specificity with greatly reduced immunogenicity in humans (Riechmann et al. (1988) Nature 332:323-327; Caron et al. (1992) J. Exp. Med 176:1191-1195).
Another approach in the production of human antibodies has been the generation of human B cell hybridomas. Applications of human B cell hybridoma-produced monoclonal antibodies have promising potential in the treatment of cancer, microbial infections, B cell immunodeficiencies associated with abnormally low antibody production, and other diseases and disorders of the immune system. Obstacles remain in the development of such human monoclonal antibodies. For example, many human tumor antigens may not be immunogenic in humans and thus it may be difficult to isolate anti-tumor antigen antibody-producing human B cells for hybridoma fusion.
For a given disease indication, one antibody isotype is likely to be greatly preferred over another. The preferred isotype may vary from one indication to the next. For example, to treat cancer it may be desirable that the binding of an antibody to a tumor cell result in killing of a tumor cell. In this case, an IgG1 antibody, which mediates both antibody-dependent cellular cytotoxicity and complement fixation, would be the antibody of choice. Alternatively, for treating an autoimmune disease, it may be important that the antibody only block binding of a ligand to a receptor and not cause cell killing. In this case, an IgG4 or IgG2 antibody would be preferred. Thus, even in a situation where a high affinity, antigen-specific, fully human antibody has been isolated, it may be desirable to re-engineer that antibody and express the new product in a different cell.
The growing use of phage display technology also points to a need for antibody engineering and expression methodologies. Phage display technology is used for producing libraries of antibody variable domains cloned into bacteria. This allows variable domains of desired specificity to be selected and manipulated in vitro. While bacteria offer a great advantage for selecting and producing antibody fragments, they are not capable of producing full-size intact antibodies in native configuration, and it is necessary to reconstitute fragments selected in bacteria into intact antibodies and express them in eucaryotic cells.